Stroke is a leading cause of death and disability in the United States, for which effective treatments remain elusive. Astrocyte survival is likely an important determinant of stroke volume and neuronal survival as loss of astrocyte viability would be predicted to augment stroke damage by glutamate release, collateral damage, and loss of their neuronal supportive functions. Thus, dissecting the molecular mechanisms by which astrocyte viability is altered during stroke would be an important advance. A principal mechanism by which astrocytes adapt to the hypoxic environment during stroke through the actions of hypoxia inducible factor 1 alpha (HIF-1a). HIF-1a, the master regulator of the cellular response to hypoxia, is a transcription factor stabilized and activated during hypoxia. HIF-1a markedly increases the expression of multiple genes including those that would be predicted to improve survival and ones that would have a pro-apoptotic function. The conditions that control the pro-survival or pro-apoptotic functions of HIF-1a remain undefined. Preliminary data suggests that HIF-1apro- apoptotic functions predominate in astrocytes during severe hypoxic stress enhancing cell death. The molecular mechanisms by which HIF-1cc induces cell death is unknown but may involve induction of pro-apoptotic members of the Bcl2 family Nip3, Nix, or Noxa. Furthermore, by associating with thepro- apoptotic transcription factor p53, HIF-1a may increase the pro-apoptotic functions of p53 during severe hypoxic stress. In contrast, pro-survival functions may be selectively induced during mild hypoxia. In fact, several reports suggest that HIF-1a plays a prominent role in the neuroprotection afforded by the mild hypoxic stimulus of hypoxic preconditioning (HPC). Furthermore, hypoxia mimetic (HM) compounds, which induce HIF-1a protein and transactivation of HIF-1a target genes, are neuroprotective when used as a preconditioning stimulus and are actively being pursued as stroke therapeutics. This proposal will evaluate the role of HIF-1a in mediating pro-survival or pro-death functions in astrocytes during stroke, following HPC, and with administration of HM compounds. We postulate that HIF-1a function in astrocytes contributes to astrocyte cell death during the severe hypoxia and ischemia present during stroke. In contrast, we predict that during the mild hypoxic stress of hypoxic preconditioning, HIF-1a function in astrocytes mediates improved astrocyte and neuronal viability. Using in vitro and in vivo methodology, we will examine these postulates. PHS 398 (Rev. 09/04) Page 107 Form Page 2 PO1 NS050315 Nedergaard, Maiken/Federoff, Howard Project 2 Response to Critiques Project 2 (former project 3) Introduction We would like to thank the committee for the careful review of our application titled "HIF-la function in astrocytes: deciphering pro-death and pro-survival roles during stroke." Since the submission of the previous application, access to new transgenic mice have strengthened the proposed experimental paradigms, recent experimental results have slightly altered the focus Aim II, and new experiments were added to Aims I & III, which strengthen the collaboration between our project and those proposed by Dr. Nedergaard. We will address these topics in order. New transgenic mice mouse models will enhance our experimental paradigms considerably. First, a transgenic mouse was provided to us through a collaboration with Dr. Frank Kirchhoff, who developed a transgenic mouse with astrocyte specific expression of a tamoxifen regulated ere recombinase (GFAP- CreERT2 mice). By breeding these mice with the HIF-laf+/f+, we will eliminate expression of HIF-la selectively in astrocytes after the mice reach adulthood, eliminating concerns of developmental abnormalities. Furthermore, we will evaluate the effect of astrocyte-specific loss of HIF-la function and its effect on target stroke volume (Aim lib), and neuroprotection provided by hypoxic preconditioning (Aim Illb). These results can then be compared to loss of HIF-la function selectively in neurons (HIF-la ::SynCre) or in neurons and astrocytes (HIF-laf+/f+::hGFAPcre) as previously proposed in the grant. Similarly to above, since the last revision of this project, we procured mice with universal expression of a tamoxifen-regulated ere recombinase [B6.Cg-Tg(cre/Esr1) 5Amc (Jax labs) [111, which will be referred to as the EsrCre mouse. By breeding these mice with the HIF-laf+/f+ mice (HIF-la ::EsrCre), we eliminate HIF-la function in astrocyte cultures. This provides us with astrocyte cultures with loss of HIF-la function (tamoxifen treated) that otherwise have the identical genotype as controls (same cultures, not treated with tamoxifen). In the last revision of this grant, we provided preliminary data suggesting that astrocytes with loss of HIF-la function are protected from oxygen glucose deprivation (OGD) compared to wildtype controls. While we replicated this finding three times prior to grant submission, subsequent experiments demonstrated no difference in OGD mediated astrocyte death with loss of HIF-la function. The original errant result was due to subtle differences in density of astrocyte cultures with loss of HIF-la function compared to controls. Now, utilizing astrocyte cultures derived from HIF-laf+/f+::EsrCre mice, we induce selective loss of HIF-la function in astrocyte cultures derived from the same embryo and plated under identical conditions removing any confounds of subtle alterations in proliferation or other effects that may alter culture density between genotypes. Interestingly, using this approach, loss of HIF-la function protects astrocytes in culture from cell death mediated by hypoxia/acidosis (see Preliminary Data). Examining hypoxia-acidosis mediated cell death was already proposed in our project, but now it will be the principal focus of Aim IIA, rather than the original focus on OGD. Finally, we added new experiments (Aim 1Cand AimlllB) to address a possible link between HIF-la and adenosine and their roles in hypoxic preconditioning. This provides a clear connection between our project and Dr. Nedergaard's proposal, which in part examines the role of adenosine derived from astrocyte-released ATP during stroke (Project 1). In Aim 1C,we explore the interesting possibility that hypoxic preconditioning alters gene expression through HIF-la and adenosine dependent molecular responses that interact to mediate neuronal protection. Recent evidence demonstrates that hypoxic stabilization of HIF-la decreases uptake of extracellular adenosine[2], and enhances conversion of extracellular ATP to adenosine[3]. Both of these HIF- la dependent responses increase the concentration of extracellular adenosine and would be predicted to enhance adenosine mediated neuroprotective effects[3]. Thus, any putative HIF-la mediated neuroprotection induced by hypoxic preconditioning could be mediated in part through adenosine. In Aim 1MB, we will test if A1 receptor antagonists alter hypoxic preconditioning utilizing the MCAO model. Responses to specific comments from the reviewers: From Reviewer: It is not clearly stated which targets will be evaluated by quantitative PCR. An interesting comparison will be made between astrocyte cultures grown in 21% O2 compared to cultures grown in 5% O2. This is relevant because brain is normally subjected to oxygen tension in the 5% range, so the baseline level of activation of HIF-1a may be different in these conditions. In addition, experiments will be performed with and without cobalt or DMOG, two hypoxia mimetics. It is unclear how much of the effect of the HM is due to stabilization of HIF-1a vs. other effects. The suggestion is that pro-apoptotic targets are more dependent on HIF-1a than some of the pro-survival targets, and some data to this effect is shown in preliminary data. A strength of the proposal is assessment of protein expression in addition to quantitation ofmRNA. 108